Academic Fraud Detection Report: Effects of Ni modified MWCNTs on the microstructural evolution and shear strength of Sn-3.0Ag-0.5Cu composite solder joints (DOI: 10.1016/j.matchar.2020.110287)

Academic fraud report · Geng Detector

Summary

This report assesses a 2020 Materials Characterization paper by Wang, Hu, and Jiang on Ni-modified MWCNT Sn-3.0Ag-0.5Cu composite solder joints. The verdict is strong concern (likely confirmed serious data-management error or fabrication) based on textual analysis alone, without pixel-level image inspection. The most damaging finding is an internal inconsistency in the isothermal aging time variable: figure captions (Fig. 7, Fig. 10) and narrative text list 240 h, while data tables (Table 2 IMC thickness; Table 3 IMC grain diameter) list 260 h. Secondary issues include (1) reported nanoindentation hardness (e.g., 0.21, 0.35, 0.36 GPa) and shear strengths (31.8, 33.2, 37 MPa) presented as single values with no standard deviation despite n=5 indentations, raising cherry-picking concerns; (2) fitted diffusion coefficients and growth exponents reported to five decimal places (e.g., 0.21349, 0.18470, 0.08957 µm/h^1/2), a spurious precision unsupported by the three-decimal measurement inputs; and (3) templated publication patterns consistent with the corresponding author's group. Limits: no image forensics were performed; the 240 vs 260 h discrepancy could in principle reflect a typographical error rather than fabrication, though the recurrence of such patterns in related papers is suspicious.

Verdict

🔴 Strong concern — Confirmed serious data-management failure / likely fabrication. The 240 h vs 260 h inconsistency between figure captions and data tables is, on its own, a fatal flaw. Secondary indicators (missing error bars, spurious precision, templated output) reinforce the concern, but pixel-level image analysis was not performed.

Key findings

Self-contradictory experimental variable (Fig. 7 / Fig. 10 vs Table 2 / Table 3): Captions and text list aging time points as 0 h, 24 h, 120 h, 240 h, and 360 h. Table 2 (IMC thickness) and Table 3 (IMC grain diameter) instead use 0 h, 24 h, 120 h, 260 h, and 360 h. The dependent variables are therefore unanchored to a defined time.
Missing dispersion statistics despite n = 5: Section 2.6 states the indenter pressed each sample five times, yet Fig. 6 and Table 1 report single nanoindentation hardness values (0.21, 0.35, 0.36 GPa) and shear strength values (31.8, 33.2, 37 MPa) with no SD, SEM, or error bars.
Spurious fit precision: Diffusion coefficients (e.g., 0.21349, 0.18470, 0.08957 µm/h^1/2) and IMC grain-growth exponents (e.g., 0.35696) are quoted to five decimal places despite Table 2 thickness inputs being only three-decimal (e.g., 3.523 µm) — inconsistent with realistic IMC metrology.
Templated-publication pattern: Reference list (Ref. 3, 5, 10, 38, 42) and authorship structure are highly similar to other 2019–2020 papers by the same group (corresponding author Xiaowu Hu) in which only the additive (Ni, Cu6Sn5, Bi, Ni-CNTs) and joint type are swapped while the DSC + nanoindentation + shear-test workflow is preserved.

Evidence highlights

Fig. 7 caption verbatim: "…aged at 150 °C for (a1) 0 h, (a2) 24 h, (a3) 120 h, (a4) 240 h, and (a5) 360 h."
Table 2 / Table 3 column headers verbatim: "0 h, 24 h, 120 h, 260 h, 360 h"
Section 2.6 verbatim: "the indenter pressed the surface of each sample for five times"
Section 3 / Fig. 6 reported values: hardness 0.21 / 0.35 / 0.36 GPa; shear strength 31.8 / 33.2 / 37 MPa — no error bars.
Fitted values in Fig. 9 / Fig. 11: D = 0.21349, 0.18470, 0.08957 µm/h^1/2; grain-growth exponent 0.35696 (five-decimal precision).
DOI: 10.1016/j.matchar.2020.110287

Notes

Image-level forensics (Western blot-style background patterns, Photoshop cloning artefacts) were not performed because only the manuscript text was provided; the conclusions above rest solely on textual and tabular logic.
The 240 h vs 260 h mismatch could, in principle, be a typographical error introduced during table preparation; however, given the missing dispersion data, the implausible fit precision, and the templated publication pattern, the most parsimonious explanation is sloppy or non-genuine data assembly.
Recommended actions: request raw nanoindentation load–displacement curves for all 5 indents per sample and the raw IMC thickness measurements supporting Tables 2 and 3; clarify which aging time is correct; if no satisfactory response, escalate to the journal editor.
This report is AI-assisted and intended for academic discussion only; final determination of misconduct rests with the publisher and institutional investigation.